classification loss
UDON: Universal Dynamic Online distillatioN for generic image representations
UDON's distillation approach is not only effective, but also very efficient, by sharing most model
Provable benefits of annealing for estimating normalizing constants Anonymous Author(s) Affiliation Address email
In fact, in a particular limit, the optimal path is arithmetic.
Facilitating Multimodal Classification via Dynamically Learning Modality Gap Y ang
Multimodal learning falls into the trap of the optimization dilemma due to the modality imbalance phenomenon, leading to unsatisfactory performance in real applications.
e8219d4c93f6c55c6b10fe6bfe997c6c-Paper.pdf
Conditional Generative Adversarial Networks (cGANs) are implicit generative models which allow to sample from class-conditional distributions.
